Key Strategies For A High Performing Data & Analytics Program (Part III)

Analytics | January 29, 2019

Delivering the Optimal Customer Experience at Scale

Why is it critical to optimize the customer journey? What impact does that make on the customer and the brand? What are the strategies you recommend to employ when optimizing this journey?

Optimizing the customer journey involves cooperation from many different teams (data, development, marketing, data privacy, IT, etc.), channels (website, mobile, email, search, etc.) and technology (DMPs, CDPs, CRMs, CMSs, etc.).

At a high level, brands know that the competitive landscape is becoming more difficult, especially for smaller, independent brands. As a result, brands need to compete on customer experience and leverage their unique relationship with their customers. For example, many customers interface with a brand over an extended amount of time, physically in-store and online. All of that rich data can be tied to a loyalty ID and stored. This can be used to recognize purchase patterns and enable targeted offers, resulting in more relevant and engaging experiences. By definition, first-party data is unique to a relationship between the consumer and brand – and is an effective differentiator in how companies serve each of their customers in ways competitors simply cannot duplicate.

To ensure the strength of a brands digital future, it’s imperative that executive leadership takes a holistic and long term view on how the organization is set up to successfully deliver on the promise of customer experience. Often times this takes a horizontal reorg around the customer journey itself. If these silos between acquisition, conversion, and retention teams are broken down, a brand can take the data they collect and then share it seamlessly. Connected technology within the stack can then provide congruent experiences for each customer across their journey. Only then can brands meet their customers where they are, right then and there, with the right experience.

The competitive landscape is changing and brands are now competing in customer experience. Laggards can no longer offer incomplete or incongruent customer experiences and hope to successfully compete for wallet share.

Analytics Techniques That Solve Problems

What are the different analytics techniques that can help solve common problems facing the marketing community? What are common problems and some potential solutions?

How can you group ‘like’ people?

Most clients are trying to make sense of the audience segments that browse their site. Behavioral segmentation through cluster analysis can help companies utilize their first-party data to better understand what affinity groups exist. Clustering is an unsupervised learning technique that leverages an algorithmic approach to deal with the vast amount of information collected. Those models surface segments in a way that drastically exceeds the capability of an analyst developing simple business classification rules. To accomplish this, take the data out of an analytics tool and use statistical programming languages to develop statistical models, then utilize decision trees to develop targeting rules to reach the audience with a personalized experience. Alternatively, some solutions now offer built-in clustering and targeting capabilities.

How can you define the customer journey?

There are a wide variety of ways a user could navigate through a company’s website and the typical question most brands ask themselves is ‘what path is most preferable?’ To answer this, specify a starting page, define the desired result and then study the sequence of events based on customers navigating the site. A Markov chain is one recommended ML model that a data scientist can deploy to pinpoint optimal steps in the journey. A personalization team can then optimize each page along the journey to better highlight the route with the highest probabilistic link to successful outcomes.

How can you incentivize purchasing behavior?

One valuable outcome-based segmentation approach is achieved through a modeling process called recency/frequency/monetary analysis, commonly referred to as RFM analysis. RFM models surface segments based on the Recency of their last purchase, the Frequency of all their purchases, and the total Monetary value of their purchases. The result allows organizations to do things like target groups who are high value, but missing in action, or highlight the biggest spenders who do not need additional incentives to convert, or identify loyal customers and supply a welcome back message.

Asking the right questions will help in identifying the challenges so best-in-class solutions can be created. In some scenarios, it is most time/cost effective to deploy ML solutions that are pre-built into martech solutions. At other times, it can be beneficial to engage a data scientist to take a deep dive and build customer models that acutely solve the business problem. Either way, the ideal solution will blend technical solutions with human creativity to deliver on the promise of providing the optimal customer experience at scale.

NOTE: This content was originally posted by Tealium in an Industry Expert Series, and can be found here.


Reid Bryant is responsible for building a world-class team focused on creating and capturing value through the sophisticated application of data science to analytics, optimization, and personalization. He brings over 13 years of expertise in quantitative fields and has led high performing optimization and analytics programs at clients like Microsoft, Barnes & Noble, Bank of America, Viacom, the Gap, Humana, Johnson & Johnson, and Chic-fil-A. He is a co-founder of the Raleigh Chapter of the Digital Analytics Association and provides digital marketing analytics lectures at the Institute for Advanced Analytics at NC State University as well as Duke University’s Fuqua School of Business.